ComfyUI Custom Extensions

An extended Load Video node with an aspect-ratio-locked, WYSIWYG crop rectangle that outputs the cropped VIDEO.

A custom node implementation for ComfyUI that integrates with venice.ai's Flux and SDXL image generation models. This project is adapted from [a/ComfyUI-FLUX-TOGETHER-API](https://github.com/BZcreativ/ComfyUI-FLUX-TOGETHER-API) to work with the venice.ai API.

Save image like the native node in ComfyUI but without metadata, supporting filename_prefix with variables like %date:yyyy-MM-dd_hh-mm-ss%, %date:yy-M-d_h-m-s% and %KSampler.seed% to include values from nodes

ComfyUI custom nodes for GGUF LLM model inference with chat support, conversation history, session management, and GPU layer offloading.

ComfyUI node suite for faster SDXL image generation on Apple Silicon with MLX.

Timed and untimed Qwen-only image and video references for ComfyUI's core MiniMax H3 Ref2VA node.

ComfyUI-Lumina-mGPT-Wrapper is a custom node for ComfyUI that integrates the advanced Lumina-mGPT model. It offers high-quality image generation with support for multiple model sizes and advanced features.

Lightweight [a/PyAV](https://pypi.org/project/av/)-powered ComfyUI nodes to load/save multimedia.

FiL Design ImageMind nodes for ComfyUI — image understanding and prompt generation (V3 API).

[a/sd-webui-pixelart](https://github.com/mrreplicart/sd-webui-pixelart) are referenced by many webui users, this node is mean to use it in ComfyUI.

Manipulation nodes for Image, Latent

ComfyUI ElevenLabs TTS Node - Text-to-speech integration with advanced voice controls

A general-purpose, dependency-free File and Folder Browser API for ComfyUI custom nodes.

Ripple Edit makes space in your workflow by moving nodes to the side as a group while keeping their relative position intact. Organize your workflow layout with ripple move and align compositions of nodes inspired by video editors. Frontend-only ComfyUI custom extension.

Native ComfyUI port of kijai's WanVideo-Wrapper TimeToMove for video generation, currently supporting LCM sampler with frame generation capabilities. (Description by CC)

Add a button that formats the workflow graph

The ultimate Image Manager & Gallery for ComfyUI. Automatically catalog generations, track lineage, and restore workflows with a single click.

NODES:Batch Image Blend, Mask Levels Adjust, Get Batch Count, Load Lora Name, Load Sampler Name, Load Scheduler Name, Load Ckpt Name.... The nodes in this repository are only used as secondary nodes.

Custom nodes for ComfyUI. Includes progressive-resolution (staged) sampling for MiniMax H3.

ComfyUI adaptation of [a/GHOST 2.0](https://github.com/ai-forever/ghost-2.0).

ComfyUI extension that adds a glowing visual effect to nodes during execution for enhanced visual feedback.

Sony Woosh audio generation custom node for generating immersive sound effects from text and video. (Description by CC)

Custom node to pause a string flow/prompt and let you edit the text

A powerful drawing and annotation tool for ComfyUI that lets you add notes, sketches, and annotations directly on your workflow canvas with full layer support.

A ComfyUI custom node implementing Colored Noise Sampling (CNS)—a plug-and-play SDE sampler that actively exploits the spectral bias of diffusion models to improve image quality.

Unified, bidirectional block-index remapping for Anima LoRAs and ControlNet-LLLite checkpoints across the 28/40/52-block architecture chain (Anima Base / 2.9B / 3.8B), applied automatically at load time -- no file conversion, no extra workflow node.

Advanced color manipulation and image processing nodes for ComfyUI: color extraction, contrast analysis, gradient generation, and image layout.

A ComfyUI custom node that randomly selects a height and width pair from a list in a config file

Advanced ComfyUI node for video generation workflows that uses multimodal LLMs to analyze start and end frames, generating high-quality video prompts with cinematic descriptions of camera movement, lighting transitions, and environmental atmosphere in English and Chinese. (Description by CC)

A node that lets you save and reuse parts of prompts (embeddings, quality keywords, and so on.)

util nodes for custom node 'comfy-mecha'

String utility for ComfyUI

Collection of custom nodes for ComfyUI designed to generate audio waveform visualizations ranging from real-time workspace previews to high-resolution images for video synthesis.

various deforum style effects

Use [a/Pilgram2](https://github.com/mgineer85/pilgram2) filters in ComfyUI

A custom node to support and load OFTv2 adapters

A set of Custom Nodes for Compositing for ComfyUI

NODES: EasyHRFix, EasyHRFix_Context, JKAnythingToString, JKBigContext, JKDynamicThresholdingMultiModel, JKEasyCheckpointLoader, JKEasyDetailer, JKEasyDetailer_Context, JKEasyKSampler_Context, JKEasyWatermark, JKInspireSchedulerAdapter, JKLilContext, JKMultiModelSamplerUnpatch, JKStringEmpty, JKStringEquals, JKStringNotEmpty, JKStringNotEquals, JKStringToSamplerAdapter

Batch samplers, schedulers, cfg, shift and steps tester custom node, automatic looping functionality for Ksampler node

Implementation of Fast Fourier Transform in COMFYUI

Utility nodes for some randomness in your workflows, like random latent sizes. A few modifications to a few existing nodes, includes nodes for sampler and model parameters. Also includes two schedulers and four samplers that can be used with any KSampler node.

Watermark removal project

ComfyUI nodes to use FluxCustomId Original repo: [a/https://github.com/damo-cv/FLUX-customID](https://github.com/damo-cv/FLUX-customID)

A ComfyUI custom node for face detection and cropping using OpenCV Haar cascades, with full ComfyUI v3 schema support and backward compatibility.

Simple resuluition calculator to convert pixel resolution and aspect ratio to laten friendlt pixel width and height size.

Nodes that support Stable Diffusion 3 Medium better.

Generate human poses from text descriptions using T2P Transformer for ControlNet/T2I-Adapter workflows

ComfyUI nodes to use [a/4D-Humans](ComfyUI nodes to use 4D-Humans)

Nodes:Generate Stable Diffsution Prompt With LLM, Translate Text With LLM, Chat With LLM

Nodes:Split Masks, Mask Selection Of Masks, Mask Region Info

This repository provides custom nodes for ComfyUI tailored for VFX and AI image/video generation workflows

Nuke-style image viewer for ComfyUI (beta): buffers, A/B compare, colour correction, video timeline, and EXR with full AOV support.

Easily download and install 2D to 3D and Flux models from Hugging Face.

Extension nodes for ComfyUI that improves automatic segmentation using bounding boxes generated by Florence 2 and segmentation from Segment Anything 2 (SAM2). Currently just an enhancement of nodes from [a/Kijai](https://github.com/kijai/ComfyUI-segment-anything-2).

Simple, no-nonsense LLM chat nodes for ComfyUI. Support OpenAI, Claude, Gemini and NoASS Roleplay.

ComfyUI nodes for Qwen-Image 2.1's official prompt enhancer, with MTP speculative decoding for faster generation.

Standalone LTX 2.3 AV director with infinite chunking, visual timeline editor, lipsync, layered audio, and per-segment prompts. No external dependencies.

Radial marking menu for ComfyUI. Alt+Click to open, hold-drag-release to select and place nodes.

FluxPseudoNegative is an advanced custom node for ComfyUI that converts negative prompts into positive ones. It's designed to enhance prompt engineering for image generation models that don't natively support negative prompts or where using negative prompts significantly increases generation time. So instead of hacking CFG we simply invert your negative words and find their antonyms!

PromptJSON is a custom node for ComfyUI that structures natural language prompts and generates prompts for external LLM nodes in image generation workflows. It aids in creating consistent, schema-based image descriptions.

ComfyUI-LTX13B-Blockswap This is a simple LTX block swap node for ComfyUI native nodes for 13B model, works by swapping upto 47 blocks to the CPU to reduce VRAM.

A small set of useful nodes which aid with the tagging process by splitting tags and strings, alternating tags from multiple sources and removing duplicates.

ComfyUI 实时性能监控插件，支持 CPU、内存、显存、磁盘和磁盘读写监控。A modern real-time CPU, RAM, VRAM, disk, and disk I/O monitor for ComfyUI.

ComfyUI implementation of '[a/Blind Image Restoration via Fast Diffusion Inversion](https://github.com/hamadichihaoui/BIRD)' Original [a/article](https://arxiv.org/abs/2405.19572)

Convert sRGB images to Ultra HDR JPEG (ISO 21496-1)

Forge fork of Anima-Artist-Mixer: a ComfyUI node pack for multi-artist mixing on the Anima model by hooking into its cross-attention layers.

A ComfyUI custom node for AI music generation using HeartMuLa models.

NVIDIA RTX Deblur & Denoise Nodes for ComfyUI\nThis extension provides GPU-accelerated deblur and denoise nodes powered by NVIDIA RTX technology via the nvvfx SDK.

A simple, quick, and dirty implementation of multiple workspaces within ComfyUI.

Load Image Media Browser for ComfyUI

The 'Eses Image Offset' node offers basic image offsetting capabilities within ComfyUI. It allows shifting image and mask content horizontally and/or vertically, with an option to wrap content around the canvas edges for a tiling effect.

UniversalLLMNode is a custom node for ComfyUI that provides a unified interface to use multiple major LLM APIs, including OpenAI, Anthropic (Claude), Google Gemini, Groq, and Mistral. It is ideal for tasks such as automatically generating high-quality SDXL prompts for image generation.

Professional image saturation control with artifact and highlight protection.

This is a ComfyUI port of PoseRetarget, originally implemented in Wan2.2-Animation. Please use it together with WanVideoWrapper.

Nodes:RR_Date_Folder_Format, RR_Image_Metadata_Overlay, RR_VideoPathMetaExtraction, RR_DisplayMetaOptions. This extension provides nodes designed to enhance the Animatediff workflow.

Forge fork of Anima-Artist-Mixer: a ComfyUI node pack for multi-artist mixing on the Anima model by hooking into its cross-attention layers.

Native ComfyUI adapter nodes for RUM FLUX.2-Klein with extra SDXL CLIP conditioning.

Extract the first and last frame from any video file or image batch in ComfyUI.

Simple Node to make panoramic images

FlexiLoRALoader - A ComfyUI custom node for dynamic LoRA weight management. Apply multiple LoRAs with flexible weight patterns and randomization features for creative AI image generation. Features: • Multiple LoRA handling (up to 3) • Weight pattern presets • Random/Sequential mode • Debug logging support

A suite of advanced ComfyUI custom nodes for a complete 'Script-to-Screen' workflow, powered by local LLMs.

ComfyUI-PromptEnhancer is a custom node for Hunyuan PromptEnhancer

ComfyUI custom nodes for WavTTS zero-shot text-to-speech with real ComfyUI memory tracking

A custom ComfyUI node for integrating with the Fal Kontext API for advanced image editing and generation.

This custom node for ComfyUI allows you to save images in multiple formats, including PNG, JPG, WebP, and ICO. [w/ComfyUI-Save-Multi-Format is renamed to SaveAs. Remove previous one and reinstall to this.]

Audio batch creation, extraction, information, resample, mono and stereo conversion. Also cut, concatenate, blend (mix) and de/normalize. Join/split channels (stereo). Signal generator (`sine`, `square`, `sawtooth`, `triangle`, `sweep`, `noise`). Musical note to frequency. Audio downloader for quick workflows which downloads its example data.

NODES:Custom Resolution Latent Node, Style Selector This Custom node offers various experimental nodes to make it easier to use ComfyUI.

A ComfyUI custom node wrapper for JoyHallo - One-Shot Audio-Driven Talking Head Generation.

Load any image straight from your clipboard into a ComfyUI workflow with a dedicated node that handles clipboard image capture automatically.

A non-destructive overlay that lights up every node by graph distance from the one you click. Diagnostic for big workflows, educational for understanding workflows others built, and surgical when you need to focus on one neighbourhood at a time.

Bake any chain of LoRAs (and any other model edits) into a single distributable .safetensors LoRA.

An implementation of Frequency-Decoupled Guidance (FDG) in pure Pytorch.

This extension adapts the [a/InstantX IP-Adapter for SD3.5-Large](https://huggingface.co/InstantX/SD3.5-Large-IP-Adapter) to work with Wan 2.1 and other UNet-based video/image models in ComfyUI. Unlike the original SD3 version (which depends on joint_blocks from MMDiT), this version performs sampling-time identity conditioning by dynamically injecting into attention layers — making it compatible with models like Wan 2.1, AnimateDiff, and other non-SD3 pipelines.

This is a custom node for ComfyUI that retrieves detailed information about an image, including its name, format (extension), DPI, dimensions, long side, short side, file size, and EXIF data. It also supports image saving

FlowRVS锛孌eforming Videos to Masks: Flow Matching for Referring Video Segmentation锛寉ou can use it in comfyUI when use this node.

GRAG-Image-Editing : Group-Relative Attention Guidance for Image Editing,you can try it in comfyUI

OmniPart: Part-Aware 3D Generation with Semantic Decoupling and Structural Cohesion,this node ,you can use it in comfyUI

Individual node snapshots and bounded link caching for smoother ComfyUI workflow navigation.

the official OpenBMB VoxCPM2 inference code as ComfyUI nodes

NODES:CXH_DeepLX_Free, CXH_DeepLX_translate